Wednesday, October 11, 2017

Cryptocurrency Incentives and Corporate Structures

The cryptocurrency world faces a tragedy of the commons and free rider issue: there is little ongoing economic incentive to contribute to an existing, major crypto project versus to launch your own token [1]. Some projects such as Tezos are helping to rectify this by thinking about developer bounties for work done, a venture fund, and other incentive mechanisms. However, the corporate structures and token distributions used by many other crypto projects today tend to exacerbate a lack of ongoing incentives. In this post I review common crypto corporate structures and brainstorm an alternative approach to try.

Existing Project Structures (AKA Where Do All The Tokens Go?)
Crypto projects tend to have the following structures[2]:
  • Individual founders or early scientists or developers. They tend to get ~10% of the initial token on a vesting schedule. (In some cases this can be as high as 20%+)
  • A company, typically an LLC or C-corporation. The company raises money, hires the initial developer staff etc. The company often ends up with anywhere from 1-2% to 15% of the tokens post-ICO.   
  • A foundation, often set up in Zug Switzerland for tax reasons. The foundation mission is typically to watch out for the long term development of the protocol. In some cases, how it is to go about this mission or how it will divvy work with the corporation is uncertain. The foundation often receives 1-10% of the tokens and may use tokens to e.g. pay developers for bounties on projects. 
  • A venture fund. Some projects have raised enough money via ICO or pre-sale that they can set up a fiat currency fund to invest in companies that use their protocol. 
The primary challenges and nuances with the model above includes:
(1) People get rich fast and may drop out. The early distribution of a large amount of highly liquid tokens to early team members can cause people to drop out of the project once they are rich enough.
(2) Most of the value escapes from the company or foundation. By distributing tokens to early team members and founders, value escapes from the company/foundation and therefore the future of the protocol. By splitting tokens between the foundation and the company, there is further dilution of effort and focus. What should the company versus foundation do? And why create multiple separate pools of tokens when you need to have a concentrated pool to incentivize ongoing contributions to the protocol? (Besides the tax efficiency of a Swiss foundation).
(3) Most of the work is done by employees of the company.  If you look at the code commits to major crypto projects outside of Bitcoin, most of the work is done by the 10-20 people that largely work for the corporation that started the cryptocurrency project. Is it possible to create incentives for these individuals to stick around as well as for the teams to scale to hundreds of people ongoing[3]? This dynamic is nothing new in the open source world. The open source projects that have tended to do best recently have companies or foundations whose focus is to monetize the protocol or project, and therefore will help pay for ongoing development (Docker, Mozilla/Firefox, Android/Google etc.).

What Can We Learn From Technology Startups? 
The standard technology company model may hold some lessons. This system, while imperfect, evolved over the last 50 years in an attempt to balance incentives and liquidity[4].
(1) Founder equity is pooled into general corporate equity. By keeping value in the company, shares accrete aggregate value that can be distributed to others. Imagine for example if the Google founders directly got their share of 30% of the profits of Google, instead of it going to the Google bottom line. Google overall as a company would be much less valuable and less able to hire and incentivize employees.
(2) There is limited liquidity for founders and employees before the company has a public (and late) liquidity event. Secondary sales by employees and founders are quite limited and are coordinated by the company in a board approved manner.
(3) Venture capital is invested stepwise based on the company hitting milestones that increase company valuation. Instead of raising $200M up front, the company will raise a seed round of a few million, a series A of $10M, and then later rounds that are much larger if the company is working. If the company does not work, it can not raise more money. This Darwinian incentive can be brutal from a founder's perspective but creates incentives to show progress and keep working hard.

The venture model clearly does not solve all the problems mentioned - for example the early employees at Google eventually left the company. However, there is an equity base that can pay for ongoing work. The difference of course is the value of a technology stock is driven by future cash flows of the company, while the value of a crypto company is often solely the value of the token held by the company.

Is it possible to combine token economics with the useful parts of the venture model?

In the crypto world, the parallel brainstorm proposal would be:
1. Put all founder, early developer, company, and foundation equity into a single entity. Lets assume for now it is the corporation, although it could be the foundation. This means 10-30% of the token resides in a single entity that can hire and incentivize employees at scale. At the founding of the project, you can raise a traditional seed round or hybrid SAFE/SAFT (which is what most crypto projects today do).
2. Give mainly equity instead of tokens. Instead of receiving a large major token distribution, most of the compensation that is distributed to founders and employees is in the form of vesting equity in the corporation. As new people join, they are issued shares backed by tokens. If their work increase the value of the token, they directly benefit from it.
3. Raise money based on equity, not tokens. You can still do token sales if you want. The company now has a non-token basis to raise money. If for example NewCoinCo holds $500 million of its own token, it should be able to raise money as a stock sale from investors at that valuation. This could be done as a crowdsale or as a more traditional financing event.

As the token rises in value, the corporation would do another fundraise via an equity sale at a higher valuation and hire more people. In other words, there is now a recurrent, milestone-based funding model for a crypto project besides selling tokens.

The money it raises via equity sales would then go to pay for ongoing development and marketing and the token base is not diluted down. One common issue with some crypto projects is their lack of experienced managers and ability to scale their teams beyond 10-20 people. As part of the equity sales, founders could ensure a subset of investors buying stock would be on the hook to help the company to hire, with marketing and strategy, and other related aspects of the business like a tech investor would.

The big drawback of course is that companies continue to generate revenue, while most crypto projects often have value solely through token value. This means that at some eventual time point you will need to sell or distribute tokens and dilute value of the company.

4. An ICO would still be held to bring capital into the company as well as create a liquid token and community incentives. By raising money in advance of the ICO, it might help create additional clarity between the equity (a security) and the token (effectively an API key for the major projects).

While there are a number of items to be worked out above, this may be a mechanism to increase the number of engineers, community managers, and marketers focused on a protocol over a longer period of time.

The challenge to this hybrid token/equity model is that in a world of nigh-infinite capital, it is hard to defer fast and early liquidity for founders contributors and investors. In some sense it is a giant marshmallow test - how to trade off short term liquidity versus building a persistant corporate entity and scaled team to drive your protocol forward for decades? As a founder of multiple companies myself, I can relate viscerally to the challenges inherent in this question.

Other Ongoing Incentives
Fred Ehrsam has a good post on how to provide ongoing incentives for protocol development including inflation, developer bounties etc. He asks the question of how do you build an evolutionary advantage into the protocol itself, so it will have incentives for it to continue to evolve.

Thinking up front about token distributions and corporate structure may be a complimentary approach to also lay a solid early foundation for the long term contribution and success of a protocol.

Choosing Financing Structures
Dan Romero and Matt Huang laid out a set of interesting comments captured below:

The "joint stock company" concept is a few centuries old. Crypto is a structural shift in many ways, so an open question is whether it should/will also change corporate structures. There might be 3 common structures that will emerge:

A. Token-only funding. Projects that require censorship-resistance as a core feature might need to fund natively via a crowd sale of tokens (as institutional capital may be scared off and there may be value in maintaining anonymity of founders). Similarly, there may be crypto projects where the userbase is inherently skeptical of any institutional involvement - which would also default to token-sales only. Finally, teams may simply decide to forgo venture capital. In all cases, there may still be best practices that emerge that signal quality efforts such as multi-year vesting via smart contracts.

Another approach other then vesting would be to have a fixed budget of tokens per year that are rewarded to the team for their efforts. Existing token holders (outside of the core team) could vote on distribution of tokens/budget per year to the team. This would be something to experiment with and is similar to some of the inflation ideas that Fred Ehrsam mentions.

B. Venture Capital (VC) + token projects. The projects will not require censorship-resistance and may accumulate value via ongoing cash flow (e.g. REP/Augur) or via token value increases. Key questions to ask before going down the token route per Dan: (i) Why do you need your own token versus just using underlying ETH? (ii) Why can't someone just fork your protocol and remove the token?

Some tokens may actually be considered securities by the SEC. In this case, a hybrid VC + token model, or VC-only model is most likely to be used.

C. VC-only financing models. VC-driven financing could occur in a number of scenarios that include non-token business models. For example, if there is no pre-sale and tokens are only available via mining or another distribution mechanism. Obviously, more traditional business models like monetizing via ads would also avoid the need for token.

There is still a lot to sort out in terms of best corporate structure and rewards for crypto projects. The most important things to keep in mind are (1) how do you create sustainable rewards/value that can be used to hire or incentivize teams over decades (2) how to create transparency on economic rewards and incentives.

Thanks to Dan RomeroMatt HuangDavid SacksLucas Ryan, and Warren Buffetcoin for comments on this post.

[1] Obviously, there are a number of other reasons to contribute to a protocol. These include wanting to make an impact, believing in the mission of the protocol, working with other brilliant people, and peer admiration and feedback. This is true of any open source project.

[2] These numbers can mean radically different things due to potential inflation (or lack thereof) of the currency, the number of rounds they plan to do (e.g. Dfinity doing a second round), and the potential to use Dutch auctions or other mechanics (a la Gnosis).  This is also all further complicated by proof of work versus proof of stake systems, as well as miner rewards.

[3] The current way for people to address this is to raise a lot of money via a pre-sale and use it to salary people. They can also give employees tokens or sell more tokens to pay for employees. However, at very large scale this model breaks down - by ongoing token sales you keep diluting the value captured by the company/foundation. If this dilution is faster then the appreciation of the token than you can have a collapse in the ability to pay a great team.

Big Banks and Blockchains
Cryptocurrency's Netscape Moment
End of Cycle?
Its M&A Time!
Market Caps and the 2% Rule
Machine Learning Startups
3 Types Of Platform Companies
Defensibility and Lock-In: Uber and Lyft
Uber And Disruption
Who Cares If Its Been Tried Before?
The Road To $5 Billion Is A Long One
How To Win As Second Mover
End Of Silicon Valley
Social Products
Hot Markets For 2015
Waiting Too Long To Go Public
What Is Your Startup Acquisition Really Worth?
5 Reasons To Sell Your Startup
M&A Ladder

Lead VC Vetos